U.S. patent number 7,961,912 [Application Number 11/811,239] was granted by the patent office on 2011-06-14 for method and apparatus for dynamic space-time imaging system.
Invention is credited to David M. Rondinone, Allan L. Sagle, Jerome R. Singer, Glen R. Stevick.
United States Patent |
7,961,912 |
Stevick , et al. |
June 14, 2011 |
Method and apparatus for dynamic space-time imaging system
Abstract
A method for creating a 3D map of the surface contours of an
object includes projecting a variety of patterns onto the object,
and imaging the patterns as they fall on the object to encode the
topographic features of the object. In one embodiment a three
dimensional image is taken in a single flash to avoid blurring due
to motion of the object. Thereafter a secondary pattern is
projected to detect changes in the initial image. The images are
processed in a computer program in a manner such that a complete 3D
map of the surface of the object is obtained in digital form.
Reiteration of the method can detect motional variation such as a
breathing human, flexure of a complex mechanical structure, or a
stress-strain testing of an airplane, vehicle, beam, bridge, or
other structure.
Inventors: |
Stevick; Glen R. (Berkeley,
CA), Rondinone; David M. (Berkeley, CA), Singer; Jerome
R. (Berkeley, CA), Sagle; Allan L. (Berkeley, CA) |
Family
ID: |
46328006 |
Appl.
No.: |
11/811,239 |
Filed: |
June 8, 2007 |
Prior Publication Data
|
|
|
|
Document
Identifier |
Publication Date |
|
US 20070299338 A1 |
Dec 27, 2007 |
|
Related U.S. Patent Documents
|
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
Issue Date |
|
|
10966095 |
Oct 14, 2004 |
7620209 |
|
|
|
Current U.S.
Class: |
382/106;
382/154 |
Current CPC
Class: |
H04N
13/254 (20180501); G06T 7/521 (20170101); G01N
21/8806 (20130101); A61B 5/113 (20130101); H04N
13/207 (20180501); G01S 17/89 (20130101); A61B
5/0064 (20130101); G01S 17/88 (20130101); A61B
5/0077 (20130101); A61B 5/1176 (20130101) |
Current International
Class: |
G06K
9/00 (20060101) |
Field of
Search: |
;382/106-108,142-154 |
References Cited
[Referenced By]
U.S. Patent Documents
Primary Examiner: Seth; Manav
Attorney, Agent or Firm: Cohen; Howard
Parent Case Text
CROSS-REFERENCE TO RELATED APPLICATIONS
This application is a continuation-in-part of application Ser. No.
10/966,095, filed Oct. 14, 2004, now U.S. Pat. No. 7,620,209 for
which priority is claimed.
Claims
The invention claimed is:
1. A method for creating a three dimensional map of the surface of
an object, including the step of: selectively projecting a sequence
of patterns on at least one portion of the surface of the object,
including an initial flash pattern and at least one secondary
pattern; digitally capturing the image of said sequence of patterns
on the object; calculating the distance of each incremental portion
of the image to the object, and storing the distance data to
comprise the three dimensional map; rendering a three dimensional
image of said at least one portion of the surface of the object
using the data from the three dimensional map: further including
the step of establishing a fiducial distance from said image to
said object and scaling said distance data according to said
fiducial distance, wherein said step of establishing the fiducial
distance includes scanning an object of known dimensions from a
plurality of different views.
2. The method of claim 1 wherein said initial flash pattern
comprises a plurality of closely spaced stripes of different
colors.
3. The method of claim 2 wherein the pattern of colored lines is
provided by a pattern or film in a projector.
4. The method of claim 2, wherein said flash pattern is formed by
an interference filter.
5. The method of claim 4, wherein said interference filter includes
a plurality of sections in closely spaced lateral array, each
section including partially reflective surfaces that establish an
interference effect unique that generates a wavelength unique to
each of said sections.
6. The method of claim 5, wherein said partially reflective
surfaces are spaced along the optical transmission axis of said
interference filter, said spacing determining said unique
wavelength.
7. The method of claim 5, further including the step of providing a
plurality of subsets of said sections, each subset generating
substantially the same wavelength, said subsets being distributed
in an orderly pattern in said interference filter, whereby colored
stripes of known spacing, color, and repetition are projected
toward the object.
8. The method of claim 2, wherein the color wavelengths of said
color stripes are matched to the maximum color sensitivities of the
device that carries out said step of digitally capturing the
image.
9. The method of claim 2 wherein the pattern of colored stripes is
provided by a digital programmable projector.
10. The method of claim 1, wherein said initial flash pattern is
used to calculate a complete initial three dimensional map of said
one portion of said object, and said at least one secondary pattern
is used to detect changes from said initial three dimensional map
and to update said initial three dimensional map.
11. The method of claim 10, wherein said initial flash pattern
comprises a plurality of closely spaced colored stripes.
12. The method of claim 11, wherein said at least one secondary
pattern comprises a plurality of closely spaced black and white
stripes.
13. The method of claim 1, further including the step of using as
said object a wound to an animal or human body, and treating and
monitoring said wound based on the three-dimensional map
thereof.
14. The method of claim 10, wherein said secondary pattern is
reiterated rapidly, whereby said updated three dimensional map
comprises a real-time record of the motion of said portion of said
object.
15. The method of claim 14, further including the step of
displaying the motion of said object as a three dimensional
movie.
16. The method of claim 15, further including the step of acquiring
the data for said three dimensional movie in real time, and
displaying said three dimensional movie in real time.
17. The method of claim 1 wherein said object is a human body, and
said three dimensional map is applied to the study of a human body
breathing.
18. The method of claim 1 wherein said object is a mechanical
structure, and said three dimensional map is applied to studying
how the strain of an object changes in time.
19. The method of claim 1 wherein said object is a human body, and
said three dimensional map is applied to the study of the human
body walking motion.
20. The method of claim 1 used for machine vision of rapidly moving
objects by maintaining a constant pattern while the objects move
through the constant pattern.
21. The method of claim 1 used to produce 3-d images for inspection
either during or after a manufacturing process.
22. The method of claim 1 used to provide computer data to develop
computer generated manufacturing procedures (CAM), or computer
aided design (CAD) procedures.
23. The method of claim 1 used with models of objects to provide
enlarged or expanded data to manufacture full scale objects.
24. The method of claim 1 used to inspect railroad rails and/or
roads for maintenance and repair determinations.
25. The method of claim 1 used to measure dimensions of human
and/or animals to determine effects of diets or of medical
treatments.
26. The method of claim 1 used to measure the dimension of injured
portions of bodies to provide the information needed to construct
prosthetic parts or devices.
27. The method of claim 1 used to obtain dental information to aid
in providing dental devices or prosthetics.
28. The method of claim 1 used to obtain building structures or
room dimensions, and/or configurations to aid in architecture for
building design and modifications.
Description
FEDERALLY SPONSORED RESEARCH
Not applicable.
SEQUENCE LISTING, ETC ON CD
Not applicable.
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to an apparatus and method for
producing a rapid time-related three dimensional image with a
numerical profile of an object. More particularly, it involves
projecting complex digitally based patterns of electromagnetic
waves (e.g. light) or scanning laser beams(s) on to the object,
photographing these patterns with a fast response digital camera,
and rapidly calculating a dimensional map of the contours, edges,
and openings of the object. In a previous patent application Ser.
No. 10/966,095, filed Oct. 14, 2004 we described the general
technique for carrying out the method. In this patent we extend
this general technique to describe several different systems. One
system uses a high intensity single flash projector(s) to obtain
un-blurred 3d images of a moving or still object. The second system
uses a series of projections to provide more information, and also
to provide time evolution information.
2. Description of Related Art
In 1905, Albert Einstein, who at that time was a patent examiner in
Zurich, developed the Special Theory of Relativity which emphasized
the importance of considering time in addition to the three
dimensions of space in describing the behavior of matter and
energy. In accordance with this concept, creating three dimensional
profiling images in a very short time period is very useful in
order to record a plethora of fundamental dynamic observations in
physics, chemistry, biology, microscopy, medicine, and engineering.
It also has a particular application to identification procedures
for security applications.
Stereoscopic photography was invented in the nineteenth century,
and has been developed since then to create very colorful
stereoscopic movies. In contrast, the development of stereoscopic
profiling with accurate detailed measurements of three dimensional
objects has been difficult to achieve. The development of digital
photography and fast computation using fast digital processing has
now provided the possibility of accurate stereoscopic imaging with
detailed dimensional measurements of the contours of an object in
real time.
Presently used techniques for non-invasive three dimensional
imaging with digital detailing of an object utilize a variety of
systems. One technique is the use of time of flight of a pulsed
laser where by the distance from the laser to the object is
determined by measuring the transit time of the laser beam. This
procedure, described by Cameron, et al in U.S. Pat. No. 5,006,721
provides fairly accurate digital topographical data. A commercial
version of the laser ranging system is manufactured by Cyrax
Technologies and several other companies. While such systems
provide good three dimensional data, they involve a quite costly
apparatus because the time of flight must be measured to a few
picoseconds, and the mirrors used to direct the laser beam as well
as the mirrors used to route the reflected beam must be exact to a
small division of a minute of arc. In addition, the scanning of a
three dimensional object with a laser beam requires a considerable
length of time, due to the fact that each incremental point on the
surface of an object must be illuminated by the beam and the time
of flight measured, resulting in a finite (and relatively long)
time for all points to be illuminated and surveyed.
Another technique for non-invasive three dimensional imaging is the
use of stereographic projections of a grid. This procedure, as
described by M. Proesmans, et al. In U.S. Pat. No. 6,910,244,
issued Jan. 21, 2003, describes the use of a projected grid for
topographic imaging. They describe a grid projection with a camera
directed to provide three dimensional imaging. The use of such
relatively static methods does not provide for the real-time
measurement of dynamic details needed for dynamically imaging and
measuring surface contour dimensions of objects which have
movement, such as a bridge or beam undergoing stresses and strains.
As stated in their patent, their application is "aimed at showing
and not measuring the object."
Applications such as rapid engineering and reverse engineering with
dynamic considerations of stress-strain relationships, measurements
of flexure of mechanical and civil structures such as airplanes,
vehicles, bridges, pipes, pipelines, steel tanks, autos and ships
require very fast imaging techniques for which this invention is
designed and applicable. In other applications as human body
imaging where, due to walking, running, throwing, swinging,
breathing and heart motion, it is important to consider the time
aspect of imaging in order to acquire realistic measurements of the
body. There is a need for such rapid imaging procedures in medical
and sports an analyses, for example, in following the progress, and
in determination of the efficacy of treatments of such diseases as
osteoporosis and skin cancers as well as other skin and body
medical and biomedical problems. There are many other industrial
applications of this invention. The position and location of parts
(of automobiles for example) on an assembly line can be measured
quickly and accurately. The invention can also be used for wheel
alignment. A major advantage of this procedure is that it provides
better angular accuracy because hundreds of thousands of
measurements are made instead of the relatively few measurements
made with the earlier laser techniques. The fast acquisition time
means images of moving wheels can be taken without blurring which
leads to better accuracy because of better feature recognition.
There are also many military and security applications from crime
and forensic scene documentation to involuntary facial scanning to
solving a jig saw puzzle: recognizing enemy assets from partial
scans taken through trees and other opaque obstacles.
BRIEF SUMMARY OF THE INVENTION
The present invention utilizes a method for dimensional measurement
of the surface contours of an object. The method employs selective
computer controlled projection of a variety of patterns onto the
object, and imaging of the patterns as they fall on the object to
encode the topographic features of the object by analyzing how the
images of the patterns projected onto the surface contours of the
object cause the imaged pattern to diverge from the projected
pattern. The object can have a motional variation such as a
breathing human, a complex flexing civil or mechanical structure,
or a stress-strain testing of an airplane, vehicle, beam, bridge,
or other structure. In the present description even faster motions
can be accommodated with no blurring. Such motions could be
associated with spinning wheels or walking patients. The images of
the object are collected by a digital camera, and are processed in
a computer program in a manner such that a complete three
dimensional map of the surface of the object is obtained in digital
form. In order to facilitate precision, the system may generate a
digital pattern which is projected onto the object in a manner so
that the pattern configuration provides the data for a computer
program to determine exact dimensions. The system may also use a
laser source of radiation by scanning the object rapidly using a
laser beam. By utilizing a rapid procedure and reiterating it, the
data provides a measure of the time variations of the size of the
object. The usual output of the digital system supplies a three
dimensional true color image as well as exact dimensional data on
the depth, width, height and contours of the surface of the object.
In one configuration, a pattern of colored lines is projected onto
the object and all of the data for the 3 D view is obtained in a
single camera shot. Advantages of the single shot are:
1) fast enough to eliminate mechanical motion effects of the
object;
2) minimizes heat;
3) can obtain 3-d images in outdoor lighting;
4) projector and camera combination are light enough to be
portable.
In another configuration the projector illuminates continuously,
and the motion of the object can be studied by measuring
differences from an initial 3d image.
The invention is particularly applicable to imaging of humans for
medical analyzes. Some examples: for full body inspection for skin
cancer wherein the size of skin malignancies can be monitored, for
breathing patterns, for reconstructive surgery, for dental analyses
of the jaw, individual teeth, and bite, for facial reconstruction,
and for gait analysis. It is applicable to design and manufacture
of prosthetic units. Likewise, the invention may be used to monitor
the changes in the body overtime to assess the efficacy of weight
loss programs, body building efforts, and the like. The system is
also applicable for body measurements for the fitting of custom
clothing and custom shoes, wherein all dimensions of the body are
acquired, and clothing patterns may be modified in every dimension
for optimal fit.
The present invention is designed to determine exact measurements
as distinguished from relative topographies. In order to carry out
such exact measurements, a range finder system and/or a standard
fiducial object can be utilized in the system so that absolute
dimensions are determined.
The invention is also useful for identification and recognition for
security applications, as it can provide motion detection and
trigger three dimensional mapping of an intruded area. Also it is
very applicable to accident analyses and prediction of mechanical
failure where dimensional changes in mechanical structures may be
analyzed in three dimensions. The invention is also very useful in
imaging the holds of ships, planes, freight trains, and warehouses,
where the images may yield numerical values for the exact amount
and distribution of space available for use, or when partially
loaded, for a determination of the space available for further use.
Another application is for security in examining the cargo space in
ships, trucks, planes, and other compartments; i.e., by determining
the volume numerically exactly so that the existence of hidden
compartments can be found. Another application is to microscopy.
Here the three dimensional sizes of the object, (for example, a
microbe), are readily determined, and the mobility of the microbe
may be measured. Recent research also indicates that weaponized
microbes in a cloud can be identified by the reflected light
spectrum and by the change in shape of a cloud. This invention
provides a significant ad vantage over the usual two dimensional
measurements used for microscopic analyzes.
In a further aspect of the invention, by utilizing a plurality of
distinct wavelengths within the electromagnetic spectrum, with the
appropriate projectors and digital cameras, infrared, ultraviolet,
or any special type radiation may be utilized to acquire and
reconstruct three dimensional images. The use of infrared is
particularly appropriate for obtaining three dimensional thermal
images which provide important information about temperature
variations in humans, animals, electronic equipment, and mechanical
structures. For such applications, the projector(s) are equipped
with emitters and the digital camera(s) are equipped with sensors
that are sensitive to infrared. Likewise ultraviolet radiation may
be used. Another application is in inspection of manufactured
products. Misalignments, displacements, or other defects are
readily detected by comparison of the three dimensionally mapped
color image with a stored image of the correctly configured
product. One such immediate use is applicable to quality control of
manufactured items, such as integrated circuit inspection. Another
immediate use is in the production of wallboard (or sheetrock), in
which each sheet must be examined for superficial imperfections as
well as surface non-planarity. Currently wallboard sheets are
examined inch by inch by a scanner system, but a single imaging
sequence of this invention at high speed may check an entire sheet
at once. This invention could also be used to measure the
dimensions and locations of parts such as doors, fenders, etc. on
automobiles as they move through the assembly line.
BRIEF DESCRIPTION OF THE DRAWING
FIG. 1 is a schematic overhead view of a layout of a single camera,
two projector imaging system designed for rapid three dimensional
dynamic imaging in accordance with the present invention. For many
applications a one projector-camera combination is sufficient.
FIG. 2 is a schematic view of the geometry of the imaging system of
the invention showing the triangulation procedure utilized to
obtain an image of the exact profiles of a three dimensional
object.
FIG. 3 is a schematic view depicting a multi-processor layout for
high speed image acquisition and conversion to three dimensional
data.
FIG. 4 is a flow chart depicting a procedure for speeding up the
dynamic imaging procedure.
FIG. 5 is a schematic overhead view of the layout of a four
projector, four camera system which can be used to obtain three
dimensional dynamic images.
FIG. 6A-6B are images of a cement mixer, which can be operated
during the imaging procedure, using the invention to image and
develop a dimensional map of the exterior surface of the mixer, and
show various dynamical stresses based on dimensional variations
observed in real-time.
FIG. 7A is an image from a scan of a basketball, and FIG. 7B is an
image of the basketball computer-generated and rotated using the
stored numerical dimensional data obtained from the scan
procedure.
FIG. 8A is an image from a scan of an upright piano, and FIG. 8B is
an enlarged image of the profile of the molding of the piano
computer-generated from the numerical data obtained in the scan
process.
FIG. 9A is an image of an automated movie doll. The image shows the
detailed set of points accumulated in the scan on the left, and
FIG. 9B is a computer rendering of the complete dimensional map
image.
FIG. 10 is a functional block diagram depicting the general method
of the present invention for generating a dimensional nap of the
outer surface of an object and applying the useful results
therefrom.
FIG. 11 is a functional block diagram depicting the method of the
present invention applied to generating a dimensional map of the
outer surface of a moving object.
FIG. 12 is a functional block diagram depicting the method of the
present invention applied to generating a dimensional map of the
human body for medical purposes.
FIG. 13 is a functional block diagram depicting the method of the
present invention applied to generating a dimensional map of the
human body and fashioning custom clothing from the map.
FIG. 14 is a functional block diagram depicting the method of the
present invention applied to generating a dimensional map of an
object and using automated machining to create from the map a copy
of the object.
FIG. 15 is schematic layout of an interference filter that projects
stripes of different colors toward an object undergoing dimensional
mapping in accordance with this invention.
DETAILED DESCRIPTION OF THE INVENTION
The present invention generally comprises a method for generating a
three dimensional map of the surface of an object, and analyzing
the dimensional data to examine the object in useful ways.
Referring to FIG. 1, the method involves providing a pair of
projectors 11, each of the projectors directed to project a pattern
of colored lines on an object 9. A single flash projector may be
composed of a flash lamp and a patterned source or a digital
projector. The flash lamp provides high intensity light for a short
time. This has two advantages: it allows the fast study of
mechanical position without image blurring and it avoids heating
the object while still providing enough light to illuminate the
object. The projectors are separated and spaced to illuminate
differing aspects of the object. A digital camera 10 is directed at
the object 9 to acquire images of the object, and the configuration
of the projected patterns of colored (or black and white) lines on
the surface of the object. The images are fed to a computer 17
which processes the images to generate a dimensional map that
numerically describes the contours, edges, openings, and surface
features of the object 9. Note that for the single flash projector,
the one-projector camera system is light enough that it can be held
in the operator's hands while taking shots from different angles or
views. Motion induced by the operator's hands is not important
because the single shot data acquisition is so fast. In some cases
it may be desired to obtain the full 3 dimensional image of the
object from different perspectives. Multiple projectors and/or
cameras can be used to obtain views that would be occluded if just
one projector camera pair were used.
By using multiple projectors 11 and the camera 10, the images
generated by the camera can readily be processed in a manner
wherein they mesh together seamlessly, and a full three dimensional
digital image of the object 9 is obtained. The apparatus may be
rotated about the object, or the object may be rotated while the
apparatus is stationary, so that all the surfaces of the object 9
may be imaged and dimensionally mapped, if desired or necessary.
The projectors and camera are all under the control of the computer
17 which sends out control signals and also receives the imaging
data which it processes to obtain a complete three dimensional
image. The seamless joining of the different views is accomplished
in software by observing that when several adjoining profiles from
the different views are found to be identical, then that is the
seam between those views, and the different views are then joined
in the display and in the numerical values of the stored profile
information.
The exact fiducial measurements of an object are obtained by
initially scanning an object with well known dimensions. The system
uses that scan as a calibration for further scans. The object is
scanned from many different angles or views. The calibration of the
fiducial measurement has the advantage that this known calibration
is the same for all views.
The computer program used to process the images of the patterns
projected on the object 9 obtained by the digital camera 10
utilizes the shapes of the projected lines to determine the
curvatures, edges, openings, and hills and valleys of the object.
The computer 17 is programmed to carry out the mathematical
function that determines the numerical profiles of the object. In
order to obtain a sharp focus of the projector images, the camera,
which is easily automatically focused, may also provide a focusing
signal for the projector focusing system thereby providing sharp
patterns to facilitate the three dimensional map determination.
The projectors 11 and the camera 10 are specified to operate using
any of several types of radiation. White light is used when true
color image construction is desired. However, in some instances,
infrared, ultraviolet, laser, or X-ray radiation are of more
importance or of more interest in imaging. The cameras and
projectors may be utilized for such procedures by simply
substituting infrared, ultraviolet, or X-ray sensitive cameras,
projectors as appropriate to the wavelength utilized.
The projectors 11 are programmed by the computer 17 through the
cable 18 to provide a series of patterns having lines extending
vertically and/or horizontally which are bent or deformed when they
are projected onto the surface of the object. The digital camera 10
receives the image of these deformations in the numerically ordered
pixels of its photosensor, the pixel data being in the form of
numbers which determine the entire topography of the object. These
ordered numbers are then transmitted via the cable 18 to the
computer 17 which processes and assembles these numbers to provide
the complete dimensionally mapped topographic image of the object.
The computation may be accelerated, if necessary, by using multiple
processors coupled together. The cost of such high speed
computation has become quite reasonable due to progress in computer
manufacturing.
A variation of the single shot system can be employed to obtain
dynamic or motional information, A single pattern is projected
continuously (or with rapid repetition) on the object 9. The
pattern image caused by the topography of the object being scanned,
which may vary in time, is then digitally acquired at a fast frame
rate which, depending upon the type of digital camera, may be 30
times per second, hundreds of times per second, thousands of times
per second, or more.
The variations of the dimensions of the object, as revealed by
successive dimensional maps taken over time, provide a time
sequence of the dynamic changes of the object.
In reconstructing the image from the calculated dimensional map, a
spline fit is utilized to smooth the digital data. This results in
a type of "best average" for the image display, and the image is
presented smoothly rather than as discrete steps common to some
digital displays.
Referring to FIG. 2, the fundamental measurement of the exact
distance from the camera receptor plane to the object utilizes a
trigonometric solution of the oblique triangle, one side of which B
is formed by the light ray B projected by the projector 11 to the
point 32 on the surface 31 of the object 9. The digital camera 10
receives reflected light from that point 32 along the direction A
which forms a second side of the triangle. The distance C which
forms the third side of the triangle is the exact distance between
the projector and the camera. The angles a and b of the triangle
are known quantities because they are established by the geometry
of the setup and the digital information supplied to the projector
and received by the camera. Therefore, using trigonometry, the
digital computer 17 calculates the exact distance A from the camera
receptor 10 to the object surface 32 using the following formula:
A=C sin a/(sin c) (1)
Since C, the distance between projector and the digital camera is
known, and angles a and c, the angles of the directions of the
projector and the digital camera are known, the calculation is
rapid and straightforward. The computer calculation is facilitated
by having the sine values in memory so that the sines of the angles
are rapidly accessed. In this invention, the values of all the
distances to the object are simultaneously calculated for all of
the pixels of the camera, each of which are, in general, different
values. This separation of all of the distances to the different
picture elements (pixels) of the digital camera occurs through the
fact that each pixel (the smallest useful image increment) provides
its individual value for the distance to the object because of the
geometry of the photosensitive receptors of the digital camera. In
other words, the individual pixels are processed by the formula
(1), and provide a complete set of distance values
simultaneously.
In order to carry out all the calculations for all of the points on
the surface rapidly, it is necessary to carry out many calculations
almost simultaneously. This is accomplished by projecting a pattern
of many lines on to the surface of the object, and using the
digital camera which, due to its numerical pixel construction,
automatically labels each individual point. All of these points are
then transmitted to the very fast computer 17 which assembles the
point calculations of distance, (from the camera to the surface of
the object), into the complete image. This procedure provides a
complete profiling of the three dimensions of the surface of the
object (a three dimensional map), including all the hills and
valleys, edges and openings of the surface of the object 9 (or
29).
In general, object surfaces are time dependent, as for example, a
stress-strain relationship or some other time dependent induced
variation or motion. To measure these time dependent effects, first
the complete image is calculated, and in the final step of that
procedure, the projector projects many fine lines at one time on
the object surface. All of the distances to every point on those
lines are accumulated in the computer memory. When the object
surface flexes or moves in any mode, the line pattern changes its
curves and distorts proportionally to the nature of the surface
change. The digital camera records all of the variations and
transmits it to the computer 17. The computer compares the new data
with the stored data. This comparison is a very fast procedure for
a computer. Using this procedure, the measurement of time dependent
changes of the object surface 31 is very fast; the limitation on
the speed of imaging time-dependent changes in the object is the
speed of available digital cameras which, at the present time, can
be considerably faster than one hundredth of a second, and even one
thousandth of a second. A continuous (or rapidly repeating)
projector that is needed for motional and time evolution studies
may be a modification of the single shot projector. For the first
set of 3D views the normal single shot pattern is used and the
regular 3D image processing is done. A second pattern consisting of
closely spaced lines is projected that has a fast repetition rate
which is synchronized to the camera with its high repetition rate.
The data for the standard flash with colored lines is acquired fast
enough to avoid blurring but it takes several seconds to process
even with fast multiple computers and this is not fast enough for
real time studies. This image from the second pattern is
sufficiently accurate and fast enough for real time studies but it
requires the 3D reference from the first pattern. For the study,
the reference is established in several seconds or less and then
data is updated every .about.10 msec or even faster.
The projector with its projected pattern can be replaced with a
programmable digital projector with a high repetition rate. This
two-pattern technique has very fast processing so that it can
provide a real time display of the 3D motion. By taking fast
repeating images with color striped patterns one can accommodate a
large range of motion.
In order to have a reference image, a mathematical procedure is
utilized which places an imaginary plane in the scene, solves for
its expected values, and uses it as the reference. This solution
both greatly decreases noise and halves the number of images, and
thus time required. Mathematical logic is included to remove the
effect of perspective inherent in cameras. The camera "sees" close
objects as larger and farther objects as smaller. Since the 3D
imaging system has recorded the distances to objects it can correct
for this inaccuracy. Because modern photo receptors have greater
resolution than current LCD projectors and other types of
projectors, the system uses the camera's increased level of detail
to augment the level of accuracy provided by the LCD or other types
of projectors alone. Logic is included which regards each resolved
line from the projector as a discrete block of many pixels and,
after analysis, smoothes these pixels intelligently using
conventional spline fitting procedures.
Referring to FIG. 3, the lines 50 projected onto the object 9 by
the projector 11 are very closely spaced and provide the basis for
a very fast imaging procedure. These are projected as a set of very
fine lines, and are perceived by the digital camera system as a
large collection of points because of the fundamental nature of the
image receptor which is divided into picture elements, (pixels).
With a digital camera system having 10 million pixels, for example,
the camera receives 10 million points of excitation each of which
has an exact geometric relationship through Formula (1) to the
angle of the lines from the projector. Each of these 10 million
pixels transmits its value directly to the computer memory 17 which
interprets each value according to formula (1) as the actual
distance from the camera photo receptor to the surface of the
object 9. The computer calculation is almost instantaneous due to
the fact that the computer memory holds all of the sine values for
immediate access. The procedure is sufficiently fast that it may be
considered as happening in real time. The transformation of the
calculated distances into the actual numerical profile values of
the three dimensional map is also very fast because that only
requires a scaling factor. The scaling factor is a simple
multiplication factor for the computer and that calculation takes
place in less than a millisecond.
The scaling factor is obtained by an initial observation of a
specific projected line which is detected by the digital camera and
then calculated through formula (1). Alternatively, the scaling
factor can be obtained by imaging an object which has an exactly
known size (a fiducial object), or by using a projected scale.
Once the entire calculation of the profile (a dimensional map of
the surface) of the object is obtained, the time variations of the
size of the object can be very readily calculated because the time
changes in the pixels of the camera change the values transmitted
to the computer memory, and these changes are immediately recorded
by the process of subtracting the values of each set of data points
from the values of the prior set of data points. In the computer,
the subtraction operations are extremely rapid by the nature of
computer design.
The computer utilized can be a very fast computer, but if extreme
speed is required it is more economical to couple a number of
computers 17, 17a, 17b, 17c, (and as many as one desires for the
selected computer speed) together as shown schematically in FIG. 3.
Each of the computers is coupled together via high speed network
cabling, and a high speed network switch within the memory
compartment 3, and the central processing modules 4. The coupling
allows all of these computers to work in unison and provide very
fast calculations. The terminal display 5 and the keyboard 6
provide a control and response display. The dotted line 7 indicates
that the system is not limited to four computers, but may be
extended to more computers if additional speed is desired.
The features regarding the speed of this system provides a basis
for dynamic imaging, which allows for the following applications in
addition to being able to very rapidly provide imaging of
non-dynamic objects.
1. Dynamic observations and measurements of deformation of aircraft
wings under stress-strain conditions.
2. Dynamic observations and measurements of deformations of bridges
and other structures under stress-strain loading conditions.
3. Dynamic observations, measurements, and detection of moving
objects for security purposes.
4. Dynamic observations and measurements of human bodies including
the aspects of breathing and movement: e.g., gait, throwing,
running, walking, swimming, swinging, and hitting (e.g. golf
swing).
5. Dynamic observations and measurements of objects which fail
under load, with determination of the failure mode.
6. Dynamic observations and measurements of humans for security
analysis and identification based on three dimensional images.
7. Dynamic motion studies with numerical analyzes with application
to human walking, human prosthetic design, human medical condition
analysis.
8. High speed quality control processing in which a product such as
wallboard or sheetrock is examined by the system of the invention,
the entirety of each sheet being examined at once in a single
imaging sequence and analyzed for surface defects or surface
non-planarity.
Other applications of the invention to non-dynamic objects (not
subject to significant movement) are:
1. Mapping of a wound in a human or animal body, so that an
appropriate dressing may be constructed quickly and accurately.
2. Mapping of a damage site in a metal, concrete or other fixed
structure, to determine their suitability to continue to perform,
or determine expected time before failure.
3. Mapping of corrosion pits in a pipeline or other critical
structure, whereby rate of corrosion (rate of deterioration) may be
observed and monitored and treated.
4. Inspection of goods during and after manufacture for quality
control, process control.
5. Mapping of objects for use in CAD/CAM production of copies;
scaling of mapped objects to scale up from model size to full size,
or vice versa.
6. Inspection of railroad rails to detect and determine maintenance
and repair situations.
7. Measurement of dimensions of human body or animal body to
determine effects of diet or of medical treatments.
8. Measure the dimensions of injured portions of bodies to aid in
constructing prosthetic parts or devices.
9. Mapping of dentition and oral cavity to aid in fabricating
dental devices or prosthetics.
10. Mapping building structures, rooms, and/or configurations to
aid in architectural design or contracting building and
modifications.
Referring to FIG. 4, a flow chart of the computation as controlled
and calculated by the computer is shown. This process is for the
real time acquisition and display of 3D motion. This motion has to
be within the field of view of the scanner. After the start 100, in
the first step 101 the projector, controlled by the computer,
projects a single flash pattern of colored stripes onto the object
which is to be scanned. The digital camera system, which is
synchronized through computer control, takes the initial image
which is processed to provide an un-blurred 3D image of the object
in step 102. This processing provides not only the initial 3D image
but also tells what plane is associated with each vertical colored
stripe. After this initial processing which can take several
seconds or less the standard color source is replaced with a black
and white source that has a large number of closely spaced vertical
lines. Each line in this pattern is assigned a plane based on the
calibration from the image data derived from the color stripes. In
step 103 the projector projects this pattern of fine lines which
the digital camera system images and transfers in step 104 to the
computer. The computer then calculates in step 105 all of the
distances from the camera to the surface of the object. This
calculation is very fast (.about.<10 msec) because it does not
have to recalibrate the plane locations for each stripe. Using
these data, the complete three dimensional geometry of the surface
of the object is obtained in step 106, and a three dimensional
rendering of the object is constructed and displayed in step 107 on
the computer screen.
It is generally likely that the object will have some movement.
Consider a breathing human, or an aircraft wing under a variable
stress as simple examples. The motion causes a change in some of
the pixels of the digital camera. A digital camera can, for
example, take 1,000 images per second. Since the movement of the
object is typically very slight in one thousandth of a second, only
a few pixels in the camera will be changed in that short a time
span. In step 108 the current image is subtracted from the previous
image to determine which pixels have changed. There are typically
only a few points that need re-calculation in the short time span.
Therefore the computer can easily do the re-calculation in much
less than a thousandth of a second. The new three dimensional
geometry in reiterated step 106 is determined very rapidly, very
nearly as fast as the object can move, since there are typically
only a few point to recalculate to update the existing 3D geometry
data. Consequently, this system provides complete motion
information about the object. In other words, three dimensional
imaging is obtained in real time for moving objects.
Because a digital camera can provide data at rates in the range of
one frame per thousandth of a second, and the computer can easily
do re-calculations for the incoming data at a rate of much less
than a microsecond, and an object generally moves very slightly in
such a short time span, three dimensional movies constructed of the
machine-rendered images are practicable.
For a small change, there is no need to recalculate and re-render
(105) the total image. Such changes are perturbations on the whole
image as defined in mathematical analyzes. Therefore these small
changes in the image can be referred by the camera back to the
existing pattern in the computer, and the computation 105 is
limited to the changes only. That procedure greatly shortens the
time needed for the display of the new image relative to
recalculating all of the image data. By this method, the variations
in such objects, for example, as aircraft wings under test, testing
of humans breathing patterns, and bridges under load can be
analyzed in real time with all of the measurements recorded in the
computer memory.
Referring to FIG. 3 again, when time permits, the system consisting
of one projector 11 and one camera 10, with or without one range
finder 8 may be utilized. In this case, when it is desired to
record and reconstruct the full 360.degree. image of the object,
the camera, projector and range finder may either be rotated about
the object, or the object may be rotated to obtain the full 360
degree view. In some cases, one may use a single camera 10 and
projector 11 and one view since one then will obtain a three
dimensional image of the region of the object that is scanned.
Referring to FIG. 5, by using four cameras 10 and four projectors
11 arrayed in gene rally Cartesian directions, the system can very
rapidly capture the front, sides, and back views of the object 9,
each of the projectors providing a computer generated pattern of
lines on the object. However, in some cases, it may be adequate and
economically advantageous to have three projectors and cameras. The
different views are all seamlessly assembled by the computer
program since each overlap of a pattern is numerically redundant
and provides the same data information. Each camera can take data
from any of the projected lines within its field of view.
FIG. 6A portrays the image of a cement mixer shown in a side view.
This image is reconstructed using all of the data assembled by
digital camera(s) and projector(s), as arrayed in FIG. 1, 3, or 5.
All of the dimensions of the cement mixer are stored, and are then
utilized to rotate and re-size the cement mixer as shown in FIG.
6B.
FIG. 7A depicts a three dimensional image of a portion of a
basketball in a front view. All of the dimensional information of
the basketball is stored in the computer, and is then utilized to
rotate and compress the image as shown in FIG. 7B. Although there
are no distinct surface features to note on the basketball, the
fact that the rendering in any rotation is smooth and spherical is
an indication of the accuracy of the method of the invention.
FIG. 8A shows a rendered image of an upright piano based on all of
the digital information about the profile of the piano determined
by the method described above and stored in the computer. FIG. 8B
is a magnified profile of a portion of the piano, showing in
profile the relief detail of the piano molding. It is apparent that
the method of the invention is capable of resolving minute surface
variations over the entire object, and rendering them clearly.
FIG. 9A depicts the rendered image of a toy based on all of the
digital information about the profile of the toy determined by the
method described above and stored in the computer. The
reconstructed facial profile of the toy shown in FIG. 9B depicts
the power of rendering from the dimensional database obtained by
the invention.
FIG. 10 depicts a flow chart showing the general method for
applying the three dimensional display and measurements of a solid
object to achieve useful results In step 200 the process of imaging
and calculating the 3 D geometry of the object. The map is compared
to an existing database of the object in step 201. In step 202 the
differences between the map and database are adjusted for a best
fit and examined, and the final step 203 the differences are
reported in a useful format. The object database used in step 201
may be obtained from prior records of design, CAD files, and the
like, or from previous dimensional maps obtained by the invention
in prior time intervals ranging from portions of a second to
months. The output of step 203 may comprise a graphic map depicting
changes overlaying an image of the object and portrayed in false
color, or contour lines overlaying the image, or other
graphic-techniques that highlight areas of dimensional change and
interest.
FIG. 11 depicts a flow chart representing the dynamic imaging of an
object or person in motion or exhibiting movement of at least some
portion, starting at step 210 with obtaining the most recent 3D
geometry map from the most recent image scan. The scan is then
compared to the previous 3D map in step 211. The differences are
then compared in step 212 and mapped into the output 214. In step
213 the differences are compared to a safety limit (dimensional
tolerances) to determine if the object is approaching or exceeding
a tolerance limit. For example, a tank undergoing pressurization
may expand normally, but it may also distort in some areas,
indicating a potential failure mode. In the example above of
quality control processing for wallboard and sheetrock, the 3D map
of the surface of each sheet may be compared to an ideal planar
surface to highlight any surface defects or surface curvature. If
the differences for any sheet exceed quality standards, the sheet
will be rejected.
Referring to FIG. 12, a method for treating the human body begins
at step 220 with imaging and calculating the 3D geometry map of the
body as described above. The 3 D geometry map of the body 220 is
compared in step 221 to a database containing the stored 3D
geometry map(s) of the body 221. The differences are then computed
and highlighted in step 222 for medical or health evaluation, and
the new image is outputted in step 223 with the differences
highlighted. This method may be used to assess changes in the body
over time for medical purposes, and also may be used to evaluate
weight loss programs, body building regimens, and the like.
Referring to FIG. 13, a method for determining sizing or custom
manufacturing for a suit, dress, shoes, or other garments, begins
at step 240 with imaging and calculating the 3D geometry map of the
body (or relevant portion of the body). The 3D geometry map is then
numerically compared in step 241 to the garment pattern dimensions.
The garment pattern is then scaled in step 242 so that every seam
will extend in a best fit to the mapped body. In step 243 the
garment fabric is then cut according to the scaled pattern and
assembled. This process produces an optimal fit to the individual
body, despite the wide range in human body types and sizes. In a
similar vein, the 3D geometry map of the body may be used instead
to select the best fitting pre-manufactured garments, based on a
complete comparison of data on existing garments with the 3D
geometry body map of the invention.
Referring to FIG. 14, a method for reverse engineering or
reproduction of an object begins at step 250 with the initial scan
that provides a 3D map of the geometry of the object. The 3D
geometry data is then utilized in step 251 to provide the machine
data to an automated milling machine to reproduce the original
object using the data of the scan. In step 252 the reproduction is
then scaled to produce a reproduction equal to the original object;
likewise, the reproduction may be scaled to any desired size
without introducing any distortion. In step 253 the automated
machine then carries out the cutting or milling to reproduce the
object.
FIG. 15 depicts one embodiment of an arrangement for projecting
colored stripes onto the object undergoing dimensional mapping. A
constructive interference filter 301 is disposed to receive
incident white light (or multi-wavelength light), processing the
incident light, and emitting stripes of spectral colored light. The
filter 301 is comprised of a plurality of sections 302 arrayed
laterally in abutting arrangement. Each section includes two
half-reflective metal surfaces 303 and 304 embedded between two
glass segments 306 and 307. A substantial fraction of the incident
white light undergoes multiple reflections back and forth between
the surfaces 303 and 304. Only light of a particular wavelength (a
particular color) undergoes constructive interference, and the
spacing of surfaces 303 and 304 of each section determines its
respective wavelength. Note that in FIG. 15 the sections producing
blue light have the smallest spacing of surfaces 303 and 304, while
section producing red light have the greatest spacing.
The sections may be divided into subsets of identical output
wavelengths, and the subsets distributed in an orderly pattern,
whereby colored stripes of known spacing and repetition are
projected toward the target object. The wavelength of each section
is given by the following relationship:
.lamda..times..times..times..times..times..times..beta..lamda..times..tim-
es..times..times..times..times..times..times..times..times..times..beta.
##EQU00001##
where
n=index of refraction of the glass;
.lamda.=wavelength of light;
d=spacing of reflective surfaces.
A significant advantage of this embodiment is that the construction
primarily of glass materials imparts a considerable heat-resistant
characteristic to the filter 301, enabling the use of high power
projection sources. Furthermore, the color wavelengths of
interference filter 301 may be selected to match the color
sensitivities of the camera image sensor(s) of cameras 10 or 11
described previously, whereby the optical resolution and the
dimensional accuracy of the invention may be maximized.
The foregoing description of the preferred embodiments of the
invention has been presented for purposes of illustration and
description. It is not intended to be exhaustive or to limit the
invention to the precise form disclosed, and many modifications and
variations are possible in light of the above teaching without
deviating from the spirit and the scope of the invention. The
embodiment described is selected to best explain the principles of
the invention and its practical application to thereby enable
others skilled in the art to best utilize the invention in various
embodiments and with various modifications as suited to the
particular purpose contemplated. It is intended that the scope of
the invention be defined by the claims appended hereto.
* * * * *